Subject: Re: [linux-audio-dev] Project: modular synth editor
From: Dave Robillard (drobilla_AT_connect.carleton.ca)
Date: Mon Jan 19 2004 - 20:56:46 EET
On Mon, 2004-01-19 at 13:39, Marcus Andersson wrote:
> Dave Robillard wrote:
>
> >On Fri, 2004-01-16 at 16:06, Marcus Andersson wrote:
> >
> >
> >>Dave Robillard wrote:
> >>...
> >>
> >>
> >>
> >>>But you're mashing the ideas of modules in an app and connecting
> >>>
> >>>multiple apps together - different problems. A jack patch bay that
> >>>works like a modular would definately be cool, but a totally different
> >>>project.
> >>>
> >>>
> >>>
> >>>
> >> From a user's point of view, why should it be different problems? I
> >>would want an application that presents both Jack clients and LADSPA
> >>plugins as modules, and let me connect them however I want. There is no
> >>reason I should have to instantiate "Jack IO modules" to access other
> >>applications. Those ports should be created implicitly by the program
> >>when I connect a LADSPA plugin and a Jack application.
> >>
> >>I understand that for efficiency, it is probably not a good idea to
> >>route all signals through jackd, but the choice of routing path should
> >>be handled automatically by the application and not by the user.
> >>
> >>
> >
> >Like I mentioned earlier, having jack used for every connection in, say,
> >a modular synth, would make your jack patching too complicated to be
> >useful. This most definatley would not be easier to use.
> >
> >
>
> Maybe I didn't express myself clear enough earlier, but I agree
> completely with what you say above. Internal signal routing should be
> handled by the application. External signal routing should use Jack.
>
> >Plugins in an app and connecting different apps is just different,
> >period. I can't think of any reason to pretend otherwise - it certainly
> >doesn't make things easier from a user perspective.
> >
> >If you want a jack audio out somewhere in your synth.... put one there.
> >You can't possibly be arguing that this is too difficult?
> >
> >
>
> Of course it isn't difficult, but if the creation of "jack audio out"
> ports can be handled automatically by the application behind the scene,
> it will simplify the user interaction. But this is not the only benefit.
> The user will also be able see the complete audio routing graph in one
> place, and don't need to switch between applications to do so.
Users can handle adding an oscillator and connecting it, what's so
different about an output? If the app does it "behing the scene"..
well, what does the user plug their "cable" into??
I think maybe you're talking about a combination analog modular and jack
patch bay? That does sound like a pretty cool idea at the start, until
you realize the things you want to route your audio through are
_seperate applications_, they're going to appear as such no matter what
you do. I really think the problem of a better jack patch bay and
writing some synth are very different problems, smashing them together
just doesn't make sense to me.
> I understand that you want to write a traditional modular synthesizer
> and that you don't consider a Jack application to be just another
> module. But it could be, and the synth could still be as efficient.
"Jack application as just another module".. now I think I see what
you're saying. I suppose it would be possible, but every application
you use would definately have to be LADCCA-fied. Otherwise this prog is
going to have to handle launching every single app you plan to use
(ugh).
> I am looking forward to use your synthesizer, whatever it will look
> like. Good luck.
Thanks, though I might actually start tackling the patch bay problem
first, as it's simpler and the UI stuff will carry over to the synth
anyway (plus I really want a modular-ish jack patch bay ;) ). Who
knows, maybe it will evolve in such a way that they are combined like
you're thinking, but I think that idea needs some _serious_ thought
beforehand.
Being a full-time computer science student doesn't exactly leave all
that much free time to hack, but I'll see...
-Dave
This archive was generated by hypermail 2b28 : Mon Jan 19 2004 - 20:58:52 EET